Re: Brain Implants & Technology Predictions [was Re: MIT scientist...]B

Bryan Moss (bryan.moss@dial.pipex.com)
Sat, 21 Aug 1999 13:11:47 +0100

Brain Implants: A Lousy Idea,
http://www.boston.com/technology/tr/august/column.shtml

There is a worthwhile argument here but unfortunately it's a single paragraph:

"[...] In humans, this balance between isolation and intercommunication is maintained by our seeing, hearing, speaking and gesturing activities, whose slow intercommunication speed, compared to thinking, most likely represents the best evolution could do to simultaneously preserve the individual and society."

He's not arguing that wiring the brain up would be impossible but that fast data tranfer would be both unfeasible and undesirable.

Michael Dertouzos replied to Bradbury's letter:

> It is precisely because I know where we are and where we are headed with
> A.I., object oriented programming, software agents, and the like that I
> make my statements. These "promises" sound great in the press and to the
> lay (in C.S.) people but they are currently not even beginning to help us
> capture and transfer the representation of human concepts, beyond simple
> electronic signals. The ability to do so in future is currently a wish.
> It may come about, of course, but we have no glimpse of a technologically
> based promise to that end.

It's clear to me after reading this that Dertouzos' motive for writing this article is because he has to deal with this sort of crap on regular basis:

http://www.media.mit.edu/~minsky/papers/Alienable%20Rights.html

> In fact I'd like to see some serious research in the area of human concept
> representation and transfer rather than grandstanding based on imaginative
> associations, which any one can make rather easily.

Robert J. Bradbury wrote:

> My general comments:
> (a) Yes, the understanding of human "concepts" is difficult, but
> should become much clearer once we have a finer map of "thoughts",
> than we currently do.

We should at least try to define the word before we map it.

> The question in my mind is whether we can have this *before* we have
> nanobots?

I doubt it.

> (b) I doubt he knows about nanotechnology, since he says "we have
> no glimpse of a technologically based promise to that end".

Wiring up all your neurons so I know what they're doing is the easy bit, doing something useful with that data is the hard bit. Harder still is putting information /into/ the brain.

> I would caution against everyone sending an email (like mine :-))
> to Dr. Dertouzos. It would be more prudent to find out from
> Dr. Minsky, whether or not Dr. Dertouzos is nanotech illiterate
> or nanotech opposed. If it the former, he can be educated and
> we should be tactful about it.

It would be interesting to hear Minsky's opinion.

> Some people may feel that my abstraction from complete knowledge
> of neuron firings to concept mapping is a stretch, if so I would
> like to know why. I view it as a simple(?) extension of current
> methods for the localization of certain "functions" to specific
> brain regions.

You can extract the raw data, you can map concepts to it, but how do you inject that data into another brain? You need a much finer undersanding to be able to move memories and thoughts from one brain to another, especially in realtime. The computer necessary to moderate interactions between two brains would probably be greater than human-equivalence, adding more brains to the collective could see a huge increase in necessary power. Brains are not made for this sort of fast direct communication. This line of reasoning only argues against direct links and knowledge injection (see the Matrix) but implants to boost speed, perhaps improve memory (of the standard kind, not the kind that can be swapped, traded, augmented, et cetera) and repair damage are likely to be common place.

BM